Curta News Letters

(since April 22, 1997)

Last Update: September 01, 2021 -- THE CURTA REFERENCE

From Erez Kaplan's Machines List server

In the process of typing a message to someone who has a missing clearing ring on a Curta, I realized that not everyone might know that the Curta handcrank should NEVER be forced backwards! (The clearing ring can go either way.) I don't know what happens if you try to back up the handcrank (except that you chew up the tip of the pawl and the ratchet teeth that try to stop you from doing so). I don't like to think about the internal damage that could result. Curt Herzstark had lots of time to think about his design, I gather, so he probably considered this aspect. The ratchet and the tip of the pawl are delicate, and don't stop a determined twist (mine is serial 33212, as i recall); later versions might be more robust.

Many handcrank calculators were designed so that you turned the handcrank backwards to subtract; someone who comes upon a Curta without its manual might try to do this.

To subtract, you pull out the handcrank and turn it the same way. For subtraction, the innards add nine's complements (except for the low-order position, which adds a ten's complement.) You can see it all in slow motion!! (Subtract zero, and think about the carries.)

Some years ago, the [true] foam rubber* pad inside my Curta's case began to deteriorate and outgas; I was worried that the possibility of sulfur compounds combined with moisture might corrode the calculator, and transferred it to a glass jar, wrapped in lots of paper towels. (Not the best substitute!) Any advice on removing the original pad? Carbon disulfide? (Catches fire really easily!!) Workmanship on the original pad was excellent, and it's not easy for an amateur to remove! I'm saving some sheet foam silicone rubber to replace it. (Sorry, don't remember where I found it.)

*Most of what we call "foam rubber" is not rubber, strictly speaking; it's polyurethane foam, which is chemically different from rubber. Materials scientists refer to such materials that spring back as "elastomers".

For lubricants, sewing-machine oil from a reliable source might not be too bad. Better would be a good grade of clock oil, for traditional weight- or spring-driven clocks. You could find out where to get it from a friendly clock/watch-repair shopkeeper. The actual source would be a wholesaler who sells clock-repair supplies and parts. Another source would be a camera-repair shop.

For Curta owners who want to see something more of the insides of their machines, but aren't mechanically inclined, you can safely remove the bottom cover, lower handgrip, and the side housing in one piece by removing the two screws on the bottom; just, please use a good screwdriver, so you don't chew up the slots.

Don't do it until you obtain a good screwdriver. (Radio Shack in the USA sells some good sets of small screwdrivers in a plastic box.) A properly-shaped screwdriver bit has parallel sides and a flat, blunt end, with edges that aren't rounded off. It does >not< look like a dull chisel!

Put the screws in a safe place (keep the cats away!), and pull gently. The housing assembly is delicate; don't drop it. There is a little key (technical term) that fits into a slot so that the cylindrical part lines up just right when you reassemble it. This key engages just before the housing assembly is completely seated; turn it a bit to seat it. The screws have fine-pitch threads, so don't over-tighten them.

You might see lots of dirt inside, and be tempted to clean it out; however, do be careful. Perhaps the easiest pitfalls to get into involve removing the setting slides and shafts. In mine, the little porous-bronze bearings at the ends of these slides are individually fitted! Keep track of them (also which is top and which is bottom) if you remove things. If you have a cat, take it to a friendly catsitter!

Also, if you have to remove a setting knob/fork from its shaft, be prepared ahead of time to deal with the spring-loaded ball inside the knob; same goes for the decimal-point markers. Losing a ball or a spring is a heartbreaking event for an amateur. The balls are likely to be a standard metric size, and the springs could be made out of the proper gauge of music wire, but not all that easily. Don't do this just out of curiosity!

Finally, does someone know whether it's practical to calculate cube roots on a Curta (other than by Newton's method, which refines a guess, and would involve noting intermediate results on scratch paper)? I have a hunch that the later Curta Model I manuals included such an algorithm (procedure), or perhaps it was in the manual for a Curta II. (Square roots are quite practical, although I'm not sure I recall all the details. Sad to say, I have misplaced my Curta, although I hope it's somewhere inside a few cubic yards/meters of dead storage. I have moved too often. :( )

Incidentally, is there any established going price for used Curtas in good condition? I bought mine new in Hong Kong for the equivalent of $80 US around 1956.

Roughly 25 years ago, Curtas were popular with sports-car rallyists, I believe because they were very well-suited for ongoing calculations (accumulated distance?) while out on the road. People who want one might have some luck advertising in sports-car (or other enthusiasts') magazines.

In my opinion, Curtas are true treasures, wonderful because they're much more uncommon than fine cameras, and because they're so beautifully designed and built.

Does someone know whether it's practical to calculate cube roots on a Curta (other than by Newton's method, which refines a guess, and would involve noting intermediate results on scratch paper)? I have a hunch that the later Curta Model I manuals included such an algorithm (procedure), or perhaps it was in the manual for a Curta II. (Square roots are quite practical, although I'm not sure I recall all the details.

For Nicholas, and anyone else without a manual, the following are the procedures provided in the manual accompanying my late model Curta Type I (model number 74043).

CALCULATION OF ROOTS

SQUARE ROOTS (Hermann's Method)

In the method to be described it is supposed, that by means of a slide rule or auxiliary tables or by judicious guessing, an approximate square root has been found. We wish to obtain a better approximation. Let N be the approximate value of R, the square root of (R squared), and denote the error in the approximation by E, so that R = N + E. The method proceeds by setting N in S.R., multiplying by N (which appears in C.R.) to produce (N squared) in R.R. The quantity 2N is then set in S.R. (without clearing R.R. or C.R.), and (R squared) is built up from (N squared) in R.R. Since

it follows that, if we neglect (E squared), E is added to C.R. Since C.R. already contained N, C.R. now reads N + E, the new approximation.

S.R. Set twice the initial approximation (i.e. 24.4) in the same positions as were used for the initial approximation before. By turning the handle in the appropriate manner in consecutive positions of the carriage, increase the number in R.R to 150, thus

This method determines as many additional correct figures as there were correct figures in the approximation. In this case the initial approximation is correct to 3 figures and the desired approximation to 6 figures. From this rule it is unnecessary to proceed further; furthermore we have in this particular case come very near to the required number (150) with 149.999.

If one requires 8 correct figures in the root and a CURTA Model ll is available, one can repeat the previous example using the initial approximation 12.25 (i.e. with 4 correct figures). One then obtains the root 12.247449 in C.R.

Remark: The rule that one gains as many correct figures in the root as one has to start with is capable of a limited number of exceptions.

CUBE ROOTS

CURTA users who have at their disposal our CURTA Tables may, by use of these tables, obtain a cube root correct to 5 figures by means of an addition and one subsequent multiplication. These tables will be forwarded free of charge on application to our headquarters.

Use of tables can be avoided by using an extension of the method for obtaining square roots described in the last section.

Let N be the approximate value of R, the cube root of (R cubed) and denote the error in the approximation by E, so that R = N + E, and (R cubed) = (N cubed) + 3(N squared) + terms in (E squared) and (E cubed).

The computation proceeds by arriving at a situation in which we have (N cubed) in R.R., 3(N squared) in S.R., and N in C.R. We then build up the contents of R.R. to (R cubed), i.e. from the above equation, adding approximately E to the contents of C.R. C.R. thus contains a quantity which approximates to N + E.

It will be appreciated that (and this remark applies equally well to the method for deriving square roots) if the derived approximation is insufficiently accurate it may be improved by repeating the process.

[Copy of message I misposted with a Cc: to "calc-list-recipients..." I assume David got his personal copy.]

David, many thanks, indeed! I truly appreciate your posting this. It's something I have wondered about for several decades, and true to the best of the 'Net, there it is, the answer. I had long thought the Curta could do cube roots. but expected an algorithm like the "fives" method, which calculates "from scratch" with no need for an approximation. However, I don't know whether such a procedure is possible.

The "fives method" is based upon the fact that the sums of the consecutive odd integers add up to the squares of the natural numbers.

1^2 = 1, 2^2=1+3, 3^2=1+3+5, 4^2=1+3+5+7, et cetera. (Of course, "^2" represents a superscripted "2".)

If you set up a calculator as for division, but increment the number in the keyboard (or setting slides) for each subtraction cycle, and do some other simple housekeeping, you can calculate a square root on this basis. However, changing the contents of a keyboard or setting slides is error-prone and a nuisance if you use consecutive odd integers.

It's much easier to enter a 5, then increment to 15, 25, 35, 45, etc.; you need to increment the number in only one column.

To backtrack a bit, you multiply the radicand by 5, and leave it in the accumulator. You then set the counter register to increment for subtraction, align things correctly, subtract a 5, and if no overdraft, change it to a 15, subtract, then 25, etc. until you have an overdraft. ("Radicand" is the number you want to take the root of.)

Five times 1^2 is 5; five times 2^2 is 5 + 15; five times 3^2 is 5 + 15 + 25, et cetera.

Please forgive me for not giving the rest of the details; I have forgotten some, and would have to reconstruct them. My Curta is either stolen or stashed away, not easily available. :( At least I can say that when you have an overdraft, you correct it without changing the keyboard/setting slides; you move the 5 to the right, and leave the digits determined so far in place. As a cross-check, I think that when the calculation is done, the counter register and the keyboard/slides should contain the same number, or should be very close. This makes sense; it's what would exist for squaring.

You might do well to practice "manual" division first, if you haven't already done so.

This "fives method" is exactly the basis for the automatic calculation of square roots in the Friden rotary electrically-driven calculators, their models SRW and SRQ; the latter also does automatic squaring from the main keyboard, and is (IMHO) a real collector's item. It was borderline troublesome in the field; more complicated! (The additional complexity of the SRW over the "4-function" STW was perhaps only something like 5% or so; it was an engineering masterpiece.)

The Friden EC-132, a pioneering and very early desktop electronic calculator (the four-function EC-130 was the world's second electronic desktop calc.) used a different algorithm in detail. in principle, it was the consecutive odd-integer method, but in detail the logic constraints (minimal complexity; it used all discrete components!) made even the fives method too cumbersome. It subtracted twice for each increment of the root digit. The second subtraction of a pair had a subtrahend one greater than the first; it was very easy to increment by one, especially because the calculator counted pulses to do its arithmetic. (It did not have any arithmetic unit in the conventional sense.)

Starting with 0, the first subtraction leaves the accumulator unchanged. Incrementing by one, the subtrahend becomes one, so the second subtraction reduces the radicand by one.

This "one" is kept, and if there was no overdraft, the one is subtracted (the first subtraction of this new pair); then it is incremented to two, and the second subtraction takes place. This second of the pair makes the pair of subtractions equivalent to a single subtraction of three from the radicand.

The "two" is kept, and a pair of subtractions, first a "2", then a "3" takes place; this is equivalent to a single subtraction of a "5".

Doubling the number of subtractions did not slow down the machine excessively; I think an all-9s root took a few seconds to calculate, but the wait was quite tolerable.

(Dave: originally, I started this message with a smaller address list; I hope you won't mind being mentioned "in the third person", particularly after you read of my enthusiasm!)

Now that I have enough memory ($5/MB, used, tested at purchase at a computer show), I can use a graphical browser. I went to Dave Hicks' (?) Museum of HP Calculators, which is a very nice site in its own right, but his section on pre-HP calculators is probably the best site on the Web; it has lots of images of mechanical calculators and adding machines, some excellent, simple mechanical illustrations of how the innards work, and an excellent description (including an image) of the Friden model SRW, the four-function + square root machine.

There's also a very detailed, excellent discussion of the Friden square root algorithm (it works well on a Curta, btw). Other Friden models not mentioned at this site are the Friden model STQ (?) (STW with automatic squaring), the SBT (STW with back transfer), and the SRQ (square root and automatic squaring; its control mechanism was so complicated that the side covers were wider than those of the other models.)

This site's URL has been hiding with an outdated URL; the "elaine-1" part is out of date; simply delete it. (Many times, I have been frustrated in trying to find the .jpg of a Friden STW; it was the outdated URL that blocked things.) I decided to try eliminating that extra "word", and all went well.

The caliber of writing and design at this site is, IMHO, excellent. I do wish that when I wrote about the Friden, Marchant, and Monroe mechanical calculators that I had known about this site; it would have made things better and easier. This site does what I would love to have done!

My text (now at Guy Ball's site) discusses the innards from a different slant compared to that at Dave Hicks' site. His site says nothing about the principles of operation of the [SCM] Marchant calculators, although my text does.

For my articles, add on friden.html, work.html, notes.html, and curta.html. to the end of this URL; e.g. ....7227/friden.html

My text is probably more personal, and less professional than that at Dave Hicks' site.

Fridens used the stepped drum mechanism (I have seem it called the "stepped reckoner" also); there were five of these; each operated two columns, I'm just about sure. The stepped drums were driven by miter gears from the main driveshaft; their shafts were parallel to the plane of the keyboard. (Not sure I didn't say this!)

A further note about the Fridens with back transfer and square root: Both of these operations required the machine to change the digits in the keyboard. The machine didn't pull down the keys like an old player piano; that wasn't necessary.

Before these operations began, the keyboard was cleared; all keys popped up so the bottoms of their keystems wouldn't interfere with the internal activity.

Normally, pushing down a key cammed a slide so that it moved a distance proportional to the digit selected; this slide positioned the gear that meshed with the stepped drum. When the machine entered its own numbers into the keyboard, a pinion drove a rack to move the slide and position the gear.

There is a basis for a cube-root algorithm; it is probably impractical for a mechanical calculator.

Concise Instructions for Curta Multiply and Divide:

For Multiplying:

Decide which of the two factors is smaller (add all the digits together "sidewise"). Enter the bigger one into the slides. (For instance, if you're multiplying 197,623 by 23, enter the 6-digit number into the slides; keep the number to the right, so that the ...23 are in the rightmost two positions.

Clear both registers, then lift and turn the top "carriage" fully clockwise. (I'm assuming a Model 1.) For this example, make three additive turns, shift the "carriage" one place CCW, then make two additive turns. OK?

I trust you know you pull the crank to do subtraction? (Turn it the same way!! NEVER force it backwards...)

Short-cut mult.: Instead of adding 8 times, subtract 2 times, shift one step CCW, then add once. What you've done is (0-2) + 10 = 8. You can extend this principle. For instance, multiplying by 99 is really easy.

For Dividing:

Enter the dividend in to the accumulator; keep it all the way to the "left". Switch the counter so it increments for subtraction (it's that oddball little two-position slide around the far side; rather stiff...

Enter the divisor into the slides, with its leftmost digit aligned with the leftmost digit in the accumulator. Subtract (more than once, probably), and watch the accumulator contents. You can be lazy, and wait until it "changes sign" (as a consequence of subtracting once too many times), or watch the decreasing number and anticipate the potential overdraft. When you can't subtract any more, shift to align the remainder with the divisor, and repeat.

Once you become proficient, you can figure out how to do short-cut division; you could develop a quotient digit of 8 by adding once (= 10), shifting one place, and subtracting twice. This is a process that probably is not practical to do mechanically.

I hope this isn't too obscure in its use of terms, and not too concise; I'm expecting you could deduce what you need to from it. I've ignored decimal points, but those tiny markers have simple rules for setting... (Each one has a very tiny spring inside, and a wee ball!)

Hello - I've corresponded with some of you already. I'm in search of a Type I Curta to go with my Type II - I've had several leads on reasonably priced ones but the leads were a bit old and the machines were sold by the time I got there. Several others are still pending - if anything turns up and it's not what I'm looking for, I'll get the information out on this list with the seller's permission.

I have a few original pieces of literature with mine, including two different types of instruction books, an original sales brochure, and a photocopy of the detailed calculating examples. I'm interested in purchasing any original literature such as this. I'm checking on an older source of original literature - if I come up with extra, I'll send a message out. I'd rather see it go to people who have Curtas and would use it to complete their collection than to see it just sit in a box somewhere.

Maybe my powers of observation are really shot due to age - but I just noticed something about my Curta for the first time several nights ago, although I've had it for several years. Sometimes I'll work a problem with the Curta while I'm watching TV just for the fun of it - I was doing this and noticed that the clearing lever is actually "detented". I noticed that as you swing the lever around, the pin drops down then pops back up - into a detent in the clearing lever. Obviously, it's to create some resistance so that you can move the lever counter-clockwise without it simply swiveling right back off the pin. Because of this, you can clear in either direction - including clearing either of the dials on the top. I suppose most all of you already knew this - but it was just another little revelation of the thought and meticulous construction that went into these fine machines.

One other thing - I've carefully removed the bottom plate on mine and pulled the lower housing off: nothing comes out and it doesn't cause any problems. I've lubricated a few, obvious things with good machine oil. Does anyone have any information about lubricating a Curta, details about how far you should *** or should NOT *** venture into disassembling one to clean and lube it, or if anyone does any of this as a service ? ? ? I don't intend to do anything more than what I've done, but I do notice that occasionally one or two of the number slides seem slightly sticky, especially just as you move from " 0 " to " 1 ".

At 08:11 AM 3/21/97 -0700, Skip Godfrey wrote:
<text deleted>
>... I was doing this and noticed that the clearing lever is actually
>"detented". I noticed that as you swing the lever around, the pin drops
>down then pops back up - into a detent in the clearing lever.
>Obviously, it's to create some resistance so that you can move the lever
>counter-clockwise without it simply swiveling right back off the pin.
>Because of this, you can clear in either direction - including clearing
>either of the dials on the top.

For those of you who are interested, here are some additional Curta details I've noticed. Contina changed the design of the Curta I and II towards the end of production. The later models had plastic cranks and clearing rings, and were supplied with a plastic carrying case. The knurled operating knob however, has always been made of metal. The plastic clearing ring no longer used the spring loaded lock pin to hold the lever in operating position. Instead they used a cheaper plastic post that the clearing lever's fork simply snapped onto when extended.

I've owned Type I (never owned a II) Curta's of both vintages, and I believe the newer style is more practical for several reasons. First the metal clearing ring will bend if the unit is dropped (horror) whereas the plastic ring tends to be more durable. Also the metal domed crank of the original design while sexier looking doesn't have the arrow that points to the counter that the flat plastic crank has. Finally the metal case that usually accompanied the old style machines had a cushion on the top made of foam rubber that tended to disintegrate and flake onto the top of the instrument, whereas the plastic case had a hard rubber cushion that seems to hold up better.

It was a few months before I gingerly opened up my Curta to see what was inside. I have been almost a mechanical technician, in all but name, since I was old enough to hold a screwdriver, and the son of a Russian mech. engineer, so the bottom screws had no especial fear for me. I guess one develops a sense for which screws one can reasonably remove, and which screws one shouldn't. Sony's electronic devices have a special symbol next to the screws you can remove without creating trouble.

(For those people whose first language is not English, "gingerly" means approximately: "carefully".)

As to opening up anything to see what's inside, that is part of one's self-education. I was distressed to see a professional educator discouraging curiosity, but that is a topic for another realm of the Internet. The proverbial cat had nine lives, after all.. Nevertheless, one should use reasonable judgment about what to dig into.

Removing the bottom (and "side") cover of a Curta is sensible provided that you pay attention to a few matters. First, use a good screwdriver. Perhaps the best that are easily available are the inexpensive sets of flat-blade screwdrivers packaged in sturdy plastic cases; try Radio Shack in the USA. A screwdriver should have a flat end that is square to the sides; the sides at the tip should be parallel, or very nearly so. DO NOT use a cheap, worn screwdriver; if the edges of the tip are rounded off where they meet the flat sides, do not use it, or carefully reshape the end with a fine sharpening stone.

The blade should fill most of the slot; it should be just a wee bit thinner and narrower. (This is basic stuff, but where else do you find two screws that are as important as these, and as hard to find replacements for?)

If you have a pet cat, be totally sure that Kitty can not get access to your work area.

Take extraordinary precautions to avoid losing those screws. I don't know about parts sources for Curtas, but finding a replacement will not be easy, believe me. A really-competent watchmaker could duplicate them (if not electroplate them) for perhaps $25 US.

Removing the cover is easy enough. (You might need to tap it with the handle of your screwdriver...) Once it's off, keep in mind that dropping the Curta in this state is a sin even greater than dropping it with the cover on! Dropping the cover is also Very Bad Manners. :)

If you are normal, you will want to learn something about how the innards work. Go to http://www.hpmuseum.org and look for Pre-HP items; look for Curta. Note that the Curta's "stepped reckoner" (sorry; that's the peculiar term I learned for it) has complement teeth; the rightmost column uses tens complement, while all other use nines complement. (Subtract zero from the accumulator to see what happens.)

Everyone should know that you NEVER turn the handcrank backwards; my Curta had a delicate ratchet and pawl to give you a hint, but careless persons could force the crank. (Probably quite a few handcrank-operated calcs. were built so that you turned the crank backwards to subtract, but not the Curta.)

When you look up inside as you subtract zero, look for the carry slides that set and reset. (As I remember, they aren't reset (the mechanical folk might say, "restored") until you turn the crank for the next operation. Mechanical one-bit memories...).

I felt confident enough to disassemble the entry-slide shafts to get access to the innards to "clean house". Thinking that Curtas were mass-produced, I very carefully saved all parts, but didn't realize that (in mine) the setting shafts were individually hand-fitted! The tiny porous-bronze bearings at the ends of the shafts were individually sized. It took me a few hours of repeated trial assembly and semi- experienced judgment of what watchmakers (I think!) call "end shake" to get things back to something acceptable. Of course, I also had to ensure that the selector gears that were positioned by the setting knobs were centered on the teeth of the central "stepped reckoner". Fools march in....

Note that the gears on the Curta have uniquely-shaped teeth; they are designed to drive in one direction only, and the driving face looks like the involute curve that properly-designed gears have.

As much as I wanted to, I have never removed my (metal) handcrank. Mine is pinned in place. It is quite likely to be a taper pin, and for the life of me, I cannot figure out which end is the smaller end! I know that one could drive the pin tighter, which is a bad idea.

With the proper information, and a small-diameter pin punch that fits (NO MAKESHIFTS!), I could drive out the pin, but only with a backing block of sufficient mass, so that the impact of the hammer would not make the whole calculator jump. The backing block (lead is probably ideal, probably at least 2 lbs/1kg. approximately.) (Yes, I know "2.2".) The crank should rest on the backing block.

There is a video tape available (or was, recently) showing a Curta being assembled at the factory; it's probably PAL, perhaps SECAM, and would need to be converted to be viewed on an NTSC (US) VCR. See Erez Kaplan's site, http://www.webcom.com/calc/

For those who have some knowledge of fine mechanisms, suitable oil and grease might well be those used for clocks and cameras.

Sorry, but the hiatus has made me forget some of the other things I had planned to include in this message...

Certainly the Curtas are among the true treasures of our civilization. I'm really glad that it was possible to manufacture such elegant mechanisms at a a profit!

Could someone tell us about what's inside a Bohn Contex? Are there any images of it on the Web? (The Contex was the "other" handheld mechanical calc. that had a significant amount of mechanism; it was not rotary, and had a 10-key serial-entry keyboard, I'm fairly sure.

One of my dreams is of a mechanical calculator that could calculate transcendental functions, and would use Curta-size parts; it would be a mechanical counterpart to the commonplace electronic handheld scientific calculators. It would probably cost $tens of millions to design and build, and would have no commercial justification (other than for those who wish to avoid use of electricity on the Sabbath!).

The innards of the Monroe PC-1421 (which I would passionately LOVE to see!!!) would be a start; that machine had a modular construction. Such a machine might best be designed upon the principles of a single-chip microprocessor, ROM, RAM, and all. It would be horribly impractical and delightful beyond measure.

When I can offer a decent sum, at some point in the future I'd love to buy a Monroe PC-1421, or at least read its service manual.

Not sure whether I still have it, but I used to own a detailed service manual for one of the STW-like Fridens, perhaps for the Model STW, which was the most popular. I'd love to get copyright permission to scan and maybe post copies of it on the 'Net. However, the whole manual would soak up the megabytes horribly; I can imagine 40 MB if in Adobe Acrobat format.

It seemed to be that these manuals had deliberate errors included that were carefully considered to be very obvious to any technician, but intended to deter copyright violators. (People unfamiliar with the machine would copy the errors faithfully.) Reminds me of an early (Intel?) semiconductor memory chip that had some errors in its layout that were bypassed instead of being redesigned out; when the Russians (no offense meant!) copied the chip, they copied the design errors as well.

On an entirely different topic: Can someone tell me about calculators (either mechanical or electronic) used in Arabic-speaking countries? In the early days of the affordable handhelds, I saw (in a NY City store) an electronic handheld with an Arabic keyboard; I greatly regret that I didn't at least ask to see its display.

I trust that most List readers are aware that Arabic-speaking countries use numerals that are quite differently-shaped from our own. If memory serves me right, their one, two, and three are easy to guess at, but a zero looks like a decimal point, and a five like a zero.

There are a couple of books in existence about the history of numerals; I found one of them (by Menninger?) to be utterly fascinating; it discussed all sorts of context.

(What follows is to some degree a repetition of material I have posted earlier, but I think with the increase in interest in the Curta it might be appropriate. I recently received a copy of the Curta service manuals.)

Folks, it's only a minor risk to open up a Curta from the bottom; there are two screws that must be loosened only with a good screwdriver that has a properly-shaped tip (square corners; not rounded; sides should be parallel or very nearly so). Wiggling the lower knurled part of the housing loosens the whole bottom housing assembly along with the cylindrical cover for the midsection. These all come off in one assembly. Put those screws in a safe place! (Keep beloved and other cats away.)

When you reinstall the housing, note that there is a "tang" that exactly aligns the cylindrical part of the housing, so the setting knobs don't drag on the sides of their slots. When the housing is almost completely repositioned, gently push it "home" while giving it a gentle twist to make the "tang" drop into its slot. You could reassemble the housing and install the screws, (even tighten them, I think), without properly aligning the cylindrical part. The Curta wouldn't fit in its case if you did, though.

In my Curta, the bottom cover, itself, is a snug fit in the lower knurled ring; it doesn't fall out, but can be pushed out. Its realignment isn't really critical, but should be done with some care to see that the screw holes are truly aligned with their corresponding holes in the frame.

Provided you use a decent screwdriver and don't lose the screws, removing this part of the housing is almost trivial. I have done it quite a few times.

Removing the setting slides and their shafts is possible, but you must keep the individual bronze bearings together with their own shafts! they are individually fitted (in mine, which I think is Ser. 33212.) You can probably interchange the shafts as long as the bearings go with them.

While you have the machine open, have a look up inside the carriage(?) to watch the carry slides. With the handcrank at home position, you can set or clear a slide, but the handcrank clears all slides before advancing the dials, I'm just about positive.

Unpinning the crank and removing the carriage(?) is another matter; IMHO, it should be done only by experienced mechanical technicians. There are delicate springs (for the carry slides, I think) that can be easily damaged, apparently. I wouldn't recommend attempting it without some experience.

After seeing the service manual, I am just about certain that you must not turn the handcrank backwards. There's a delicate ratchet inside it to suggest that to you, but until now I didn't know whether it would do harm. It probably will, and more than just the ratchet teeth will be damaged.

(Don't assume that what is true for some other, much bigger handcranked calcs (such as the Monroes, and probably Odhner and such) holds true for Curtas; Curta subtracts by adding complements, not by running backwards as do many other machines.) (For anyone who doesn't know, pull out the handcrank to shift to subtraction mode, and turn it the same way.)

I bought my Curta in Hong Kong while in the Navy, around 1956 (for $80 US!). In succeeding decades, it seems that the sponge rubber in the bottom of the case started to decompose, and I suspected that the gases evolved might be compounds of sulfur. I moved the calc., wrapped in paper towels, into a well-cleaned apple sauce jar, because I was concerned that the gases would corrode the metal in the calc. Can anyone shed light on this?

Btw, almost all of what most people call "sponge rubber" is not rubber, but polyurethane, a chemically-different elastomer. (Its combustion products, I understand, are extremely toxic, being closely related to cyanide.)

I bought my Curta I in Hong Kong (for the equivalent of $80 US!) about 1956; my serial no. is, as I recall, 33212.

Some details: Metal clearing ring and crank, gray sponge rubber disc in the bottom of the case (It's not sponge polyurethane, which is also an elastomer, but different); case seal is a white O-ring. Bronze bearings for the setting shafts were (in effect) hand-fitted and not interchangeable (Guess how I know!).

It's good to realize that what we loosely refer to as "foam rubber" is only rarely rubber. It's almost always foam polyurethane, which deteriorates into a nasty sticky mess after many years. Foam rubber becomes stiff, and eventually rigid; it may form crumbs, but not sticky messes.

The foam rubber disc in the case began outgassing (sulfurous odor) years ago, and I worried that the gas and moisture might combine to form an acid that would corrode the innards. So, years ago, I put the Curta itself into some dry paper towels for cushioning (less likelihood of acid!), a polyurethane bag, and then into an apple sauce jar, and fastened the case to the jar with something like good string. Really sad to say, as precious as it always has been, I don't know any more whether I still own it; it is, I hope, stashed away in a rather-inaccessible storeroom. It could have been discarded, or stolen by others, probably in the course of several rather-frantic moves.

> It's good to realize that what we loosely refer to as "foam rubber" is
> only rarely rubber. It's almost always foam polyurethane, which
> deteriorates into a nasty sticky mess after many years.

The biggest calculator-related problem I've had with polyurethane is the cushion pads that Canon used under the keys of their desktop electronic calculators. They turn to goo and stick the keys down!

Some time ago my ancient Hewlett-Packard frequency counter quit working. It was 1960s vintage, with a row of discrete-transistor decade counters and incandescent-lamp displays plugged in to a motherboard in the bottom of the case. When I opened it the problem was immediately obvious - one of the transistors on the top of one of the counter boards had fallen off its legs! There were 3 legs still soldered to the board, and the metal transistor can lying in the bottom of the instrument. I prodded the next board, and the transistor in the same location fell off too! And so did the next one. There was nothing else different or unusual about them - the same devices were used all over the boards. Those in other locations all seemed perfectly solid, but those in this particular spot all seemed to have developed a most peculiar ailment.

About this stage I noticed a strip of partly-decomposed polyurethane foam about an inch wide glued to the underside of the lid, presumably to stop the lid (or the boards) from rattling. It was immediately above the row of "rotten" transistors! I replaced the foam with felt, and replaced the (germanium) transistors, and the instrument is good as new again.

I don't know what the corrosive ingredient is in polyurethane, but it certainly has no place in with delicate equipment.

Although some people don't like to even go near gun shops, nevertheless some of what they sell can be useful for other purposes. Their ear defenders (for noise) can often be inexpensive and good. Gun cases can be very good for some musical instruments (such as recorders (block-flutes)) sold without cases.

Tools for working on guns are likely to be quite good, and some, such as screwdrivers, are rather well adapted to working on calculators. Gun oils, greases, and other lubricants are likely to be of high quality. (But, if you can find out, use something close or identical to what the factory service people use(d). Of course, such information is generally quite hard to get, if not impossible.)

Although the end uses are very different, both guns and mechanical calculators are mechanisms that really have to be very reliable.

Hello all - I was contacted by a person who removed the "decimal pointers" on a Type II Curta in order to clean and lube them. Both the upper and lower tracks have a screw that serves to block an access hole - remove the screw, and the pointers will slide right out.

The problem is that each pointer has a tiny steel ball (only 1.0 mm in diameter) and an equally tiny spring behind the ball inside the pointer housing. Some of the balls were lost - and 1 or 2 of the springs. I managed to find the steel balls - you can't imagine how tiny they really are - and sent a few plus kept a supply just to have them. But I can't locate springs. They have to be approximately the same diameter as the balls or slightly less, and about 2.0 mm or 3.0 mm in length. The smallest diameter I've been able to locate is about 1.44 mm and they simply won't fit.

Also, re-inserting them looks to be tricky - you apparently have to either devise some sort of tool to compress the ball and spring into the pointer housing, place it into the "track" right by the access hole, and then slide the pointer assembly out of the tool and into the track. Or - with wood, hard wax, etc., you need to plug the access hole so that it's level with the track, insert the pointer assemblies, then remove whatever was used and re-insert the screw to block the access hole.

When I bought a Mercedes Euklid many months ago, it was shipped from Germany - and packed with extreme care. Regardless, just the shock of someone's "gentle handling" was enough for the operating handle to break off the mounting hub. My friend who found it for me was really disappointed - the machine was, and is, in wonderful condition and the crank was perfect which is a bit rare for Euklids.

I took the hub and the handle to a local gunsmith. He re-attached the handle to the hub, using 1 or 2 pins and some form of solder. You can see the fracture line but it's not really all that obvious - and the machine works perfectly. I believe the cost was $40, which I thought was quite reasonable. Had I wanted to spend more money, I'm sure it could have been filed, buffed, possibly re-plated, and the fracture line made almost invisible. But that wasn't my goal - and these things should be disclosed, anyway, if a machine is ever sold.

So - if you have a problem with a calculator and it doesn't necessarily require that parts be acquired, it might be worth checking with a gunsmith. In my experience, they are extremely capable craftsmen - and have a lot of pride in their work, plus a well-developed sense of curiosity.

I'll jump in with $.03 worth. Has anyone seen the videotape which shows assembly of the Curta?

Making springs isn't out of the question; you should have a clock repairer for a friend! Try music wire; hope the required gauge is available! The mandrel will need to be undersize, to allow for the natural unwinding. Try a clock/watch repair supply house. These would be simple compression springs; you might be lucky. Try also: PIC Design, ?Sterling Instrument/Designatronics?, or Small Parts. All come from memory.

The Thomas Catalog (many green-covered volumes) is a wonderful place to look for tiny steel balls. Who's saving wide-stroke ball pens? (Just be sure the balls are shiny! You don't want a textured-carbide ball!)

If by any chance you need to work on a Friden carriage, there's a similar risk; at least the springs and balls are larger. This is "major surgery", not often needed. It was called "splitting the carriage", and you needed a good handful of rubber bands. The 20 accumulator dials, at least, must be held in place against the correct half of the frame before you split it apart. (I know I've written about this before.)

I was lucky; not sure whether I removed these markers, but when I first got acquainted with my Curta, I was delighted to see such a refinement. Probably only the German Diehl desktop machines would contain such refinements. (Their nonprinting desktop model(s) reminded one of Fridens; they were exquisitely built. Think Bugatti? There's a bit more info on Diehls at Mr. Bl�mich's German site, recently noted in this List. I really hope to see another Diehl before my time comes!)

As to blocking off the hole for the retainer screw: Do you really have to? If so, try to find a screw with the right threads (clock and watch folks, again; see the parts list for possible thread size info.) and either choose one with a really-tiny head, or modify the screw. Seems that it shouldn't be a problem to assemble these; they were mass-produced, after all.

I seem to recall assembling these markers by trying to keep them pushed to one side of the hole, as much as the geometry would permit. However, do not trust this possible recollection.

Have a good look at some watchmakers' benches. Some have "aprons" to catch dropped parts. As well, be really glad you didn't lose a part like the little sliding 5-tooth selector gears!! Note their unique tooth profile, which was used at least part of the time for Contina's trademark. Because they are driven always in one direction only, the back side of each tooth is a simple "ramp".

Losing even a screw from a Curta would be a Very Sad Event, believe me! You absolutely do not buy these even at the best industrial hardware stores. Think clockmakers and tens of bucks... Many, perhaps even all, screws are custom-made, with metric threads, naturally.

Fwiw, IBM had in-house *incompatible* screw-thread standards for its business machines in the 1930s, probably earlier, and maybe later, just about guaranteed!

Happy thought, Skip! Makes a lot of sense. I have a hunch that clock and camera repair people would also be good to make friends with.

(Will we ever see a calculator like the innards of a modern camera, with electronics all scattered about among the machinery? I doubt it. However, I do recall a thoroughly-amazing printer mechanism in a low-cost handheld calc. that had a drive motor, one electromagnet, two optical-interrupter timing wheels, and a mechanism that was so baffling I finally felt I'd met my match. By genius-level integration of the electronics with the machinery, that little feller did all printing tasks, even color changing. It was *very* humbling. I didn't have occasion to really study it, however.)

Thanks to Ray Mackay for pointing out that if you use a lathe to make springs, wear safety glasses! Also, naturally, set your spindle (headstock) for *really* low speeds. You might do better to disengage the drive and turn the spindle by hand. However, if you do so, be really sure the motor won't try to turn the spindle. Use at least two methods of disabling it. Allow for the possibility that the spring will try to unwind itself.

If you're making springs larger than those for the Curta decimal markers, you could wind a much longer length and cut your long winding into the needed lengths. I think I've seen spring stock that you could cut.

Hello all - I noticed on both a Type I and a Type II CURTA that there seemed to be a "drag" when I turned the operating crank. I was afraid that they both needed major cleaning. With a little careful inspection, it became obvious that the small knurled knob was dragging a lot, almost binding - and that the main shaft seemed to turn freely. I know many people are reluctant to even go near their CURTA with a screwdriver, and in general that's probably a good idea - but what I did to fix the problem, and the difference it made, is worth mentioning.

On both machines, the knurled knob is held in place with a single screw - it's a small screw head and a thin screw slot so it requires a fine screwdriver. If you loosen the screw a bit, the knurled knob will pull straight off the pin that it attaches to. When I removed the knob, I was amazed at the amount of dried black "gunk". I cleaned the shaft and the groove for the screw with good Starrett instrument oil, a cloth, and a toothpick. Then I oiled the shaft liberally, slid the knob back on, and worked it back and forth quite a bit. Then I re-cleaned everything, including using the small screwdriver to force an oily cloth down into the recess in the knob. Finally, after drying everything I put a very small amount of on the shaft, replaced the knob and worked it back and forth a few times, then gently tightened the securing screw.

The difference is feel is really substantial - if your CURTA hasn't been cleaned and oiled for a long time, it may be worth trying. I really don't see how you can hurt anything if you use a good-quality screwdriver and are cautious in using it.

OK Curta freeks, For thirty years I've had good luck using the following technique. Not hi-tech but it works! Obviously, watch out for fire and wear protective gloves.

This makes all the difference in the world and returns that silky, as new, feel. You can tell by the sound if everything is properly cleaned. It makes nice clickity sounds! I love my curtas and over the years have learned to really make them hum. The best part is that you can operate them by feel, in the dark. The hardest part is reading the numbers while flying down bumpy roads at speed in the middle of the night!

I tried it just last night, with my roughest Curta. I took it back out of the Coleman fuel undt (or is it "mit" - just my attempt at 5 cents worth of German) oil this morning b/4 I left for work - it's drying out today. There WAS a fair amount of dirt particles and general "gunk" in the can - not a huge amount, but it was very obvious. I'm not sure that I had quite enough oil in the Coleman fuel - the oil I used was Starrett instrument oil but I also bought some sewing machine oil. If it seems a bit "dry" afterward, I may do it again but with a higher concentration of oil. I will probably still double-check the roller knob and oil it lightly - I don't think it's a bad idea even if you do use the Coleman Maneuver first.

I just bought a Type II which is in very good condition - but a bit in need of a cleaning. I've found that using a Q-tip and a bit of standard, drugstore alcohol can do wonders. I went over the entire surface, the faces of all the digits, and basically tried to gently remove any accumulated gunk. I found that the wheels in the results register and rotation counter can be easily rotated - I used the cardboard "stick" of a Q-tip, with one end broken off. It was easy to set all the dials at the same number, clean each one, then click them to the next number. There were a couple which didn't clean as well as I would have liked, but both the end result and the pile of dirty Q-tips showed that it accomplished a lot.

One other thing. Thanks to Jim McDermaid, I found out that there is something called a "lacquer stick". It looks like a big crayon - but it's a paint compound of some sort. His source - and the one I used - is a company that sells supplies for re-conditioning old radios. In that context, the sticks are used to re-finish the numbers in the dials of old radios. I haven't used mine yet - but the instructions are to clean out the paint in the recessed number, rub new paint in with the lacquer stick, wipe off any excess, then let it dry. They're available in white, black, gold and red.

If anyone is interested in attempting this in the course of going through a machine, please let me know - it's not appropriate to list commercial sources on this list but I'll be happy to email the information to anyone who's interested.

I have used these on a few occasions. They seem to be the sort of item that can be found at industrial-supply houses who have good variety. Ames Supply used to carry them, I think; a quick peek at the Web implies that they might not carry calc. repair tools any longer, but it could be worth a check. (In the 1960s, Ames was one of the best sources of specialized typewiter (and therefore, calculator) repair tools. (Ernie? Any comment?)) Any recessed markings can be refurbished by these lacquer sticks; just about sure they come in a limited range of colors.

It seems that the vehicle ((?) the gunk that carries the pigment) dries to a skin that prevents further curing deeply into the body of the stick; don't really know.

Try an artists' supply store that has a lot of variety, and ask! Tell them what you need, and you might be lucky, as well as have a great choice of colors. Imho, these sticks should be part of any restorer's tool kits.

Btw, in general, I'm *not* an expert, just over 60; there's a difference! I love to write, as you know. Occasionally, I've gone owerboard and made a fool of myself when I see the comments of a real expert. This is a general remark; caveat lector! Nevertheless, the content of this message should be solid. (Better than the inside of alacquer stick, I hope.)

Some time, one of us will describe double-shot injection molding, used for the Friden EC-130 keys and the large-diameter number/letter plates that surrounded some later rotary phone dials. I really ought to go...

Hello all - one of these days my curiosity is going to get me in deep water, but so far so good. I do try to be very, very cautious about taking even the slightest additional step. But one thing I just completed, on a bargain Curta that I got, might be worth mentioning.

The gray Type II is very, very nice cosmetically - but the sliders were very sticky, and it seemed like one or two got worse, not better, as I tried to work with them. So, I removed the bottom two screws, the base plate, and pulled off the barrel. When I looked at the serpentine shafts that the slider knobs are one, it seemed like I could actually see a clear coating - I ran a fingernail down the shaft of one of the stickiest knobs and was able to pull off something that seemed almost like parafin (like you use to seal some kinds of home canning).

Long story short - I drug out the trusty Q-tips and alcohol, cleaned the shafts all over (and got lots and lots of off-color gunk from them), ran the sliders back and forth, put a moderate amount of the good Swiss clock oil on them, again worked the sliders, then lightly rubbed the shafts with clean Q-tips to remove excess oil.

The difference is absolutely amazing. Thought it was worth passing on. The only minor problem in any of this is getting the top of the barrel to seat into the upper housing (squeeze selectively, all the way around, until everything finally drops in) and getting the two screws in the base plate to drop properly into the threaded holes. If a person is cautious, I can't see how you could cause any damage.

Sounds simply exemplary; no less. Some of the budding calculator collectors seem to have the right instincts; you surely do. (Rick Bensene is another.)

My Curta started to become grungy inside, although not too bad. I went the next step, and removed the setting shafts; I removed the little retainer plates (two, iirc) and then the little porous-bronze bearings at the bottom of the setting shafts. I took precautions not to lose any tiny parts.

Once the cleaning was done, I reassembled everything, and tried moving the setting knobs. Egad!! Some were tight, and others had lots of "end shake", to use a clockmaker's term. The bearings were selected, and hand-fitted! It took a lot of trial-and-error swapping to restore things to normal.

Apparently, for some odd reason, the lengths of the setting shafts had loose tolerances, and they weren't freely interchangeable. Mine is (or was) Serial 33212, fairly sure; that seems to be one of the early ones.

This was quite a surprise; interchangeable parts have been around for well over a century, and in such a beautiful mechanism, I would have expected interchangeability. (It's just possible that the machine tool that made the shafts couldn't hold tight tolerances on length; still seems odd.)

Btw, I wouldn't recommend more disassembly. On at least some Curtas, it's far from obvious which end of the handcrank pin is the smaller. (Almost certain that the service manual makes this clear.) Once you remove the handcrank, you're getting into non-obvious territory; try to get the service manual if possible. This is not for talented amateurs, I'd say. (I probably wouldn't attempt it, definitely not without the manual.) Furthermore, if you remove the [carriage] (which carries the dials and clearing lever), there are very-delicate springs (for the carry slides, iirc) that the service manual warns about; apparently, they are quite easily damaged.

Well, being ever inquisitive, I decided to go one slight step further in cleaning and lubing this newest Curta - what's that music I hear in the background, something about "fools rush in . . . where Angels fear to tread" ? ? ? Really, I'm glad that I did - the improvement was not earthshaking but it was noticeable.

Combining what I'd observed in the "bottom end" of the Curta, what I'd looked at in the Curta repair manual, and comments from Nicholas Bodley about removing the hand-fitted bushings in the setting shafts, I decided to see about lubing them.

*** ONE CAUTION *** I didn't want to remove the bushings; I couldn't see any reason to and I definitely didn't want to get into the problem with matching the bushings to their shafts which Nicholas mentioned. I had read this somewhere before, that the bushings were hand-fitted to the shafts during manufacture. I doubted that the bushings would simply fall out but wasn't sure, so - I KEPT THE CURTA UPSIDE DOWN THE ENTIRE TIME I WAS LUBING THE BUSHINGS. Maybe it's unnecessary but I felt - and still feel - that caution is indicated here.

There are 3 semi-circular plates on the bottom of the Curta housing - they're directly under the lower ends of the setting shafts. Each plate has two screws holding it to the housing - the screwholes in the plates are "keyholes" so you only have to loosen the screws and the plate will slide away. Perhaps you can adequately lube the shafts from above, without removing the plates and exposing the bushings and tips of the shafts - but I chose to remove the plates. I only removed 1 plate at a time. Then, I used the finest blade of my jeweler's screwdrivers, put a moderate drop of the good Swiss clock oil on the tip of each shaft, ran a Q-tip along the tips of the shafts to pick up any excess oil, then replaced the plate. One of the three plates has a little notch on one end to clear an adjacent nut.

After this, I decided to put a drop of oil on the central shaft where it fits into the bottom of the main housing. If you look down into the center of the Curta and move the operating crank up and down (shifting between addition and subtraction), you'll see where the shaft moves in the housing. I also put a tiny bit of oil on the circular disk with an indent on the bottom (this, and the spring-loaded follower, provide the "home position" that you feel after you've made 1 complete revolution), as well as the spring-loaded pawl that's intended to prevent backward rotation.

Several people have emailed wondering in a nice way if it's a good idea to mention all of this - I guess the fear is that someone may be tempted to jump in over their head and possibly really damage a nice Curta. Perhaps that could happen - but I'm not that gifted at mechanics and I haven't damaged anything yet. I would not encourage anyone to attempt even removing the base plate if they are not comfortable using basic tools - and certainly many people are not. But I know that these few little attempts have made a definite improvement in several Curtas of mine - especially cleaning and lubing the sticking sliders on the setting shafts and also the knob on the operating crank, something I've mentioned in the past. Besides eliminating "sticking" and drag, there is a more solid feel when the Curta is lubricated - I think the oil simply cushions things and increases the feeling of "smoothness", something that is completely different from "not sticking".

First, about using alcohol for cleaning the surface of Curtas. I have never experienced any problems with this - but I have always used what I would call standard "rubbing alcohol": the bottle I have is 70 % alcohol. I was told that someone had used alcohol for cleaning a Curta and felt that some of the finish actually was removed - although I haven't had this happen, I suppose it is possible. I wonder if it might have been the result of using pure alcohol however. I have a small container of 99 % alcohol at work - it was packaged with a cleaning kit for computer tape drives. Using 70 % alcohol has not caused any problems that I am aware of.

Second, does anyone have a few stylii for Addometers, Lightnings, etc. that they would care to sell or trade ? If so, please drop me an email.

I'll jump in with some speculation on this. Assuming the Curta's exterior is aluminum alloy, it is probably anodized. However, anodizing, by itself, is essentially colorless, but (apparently) rather porous (at least initially). The black is very likely to be a high-quality dye, I'd think. It's possible that some dyes would be alcohol-soluble. However, removing any substantial amount of the dye by wiping with alcohol seems unlikely.

The Metalphoto (tm) photosensitive plates are anodized Al alloy. The pores are filled with photo. emulsion, which is exposed and developed by conventional means. Once wet processing is complete, the plates are boiled (iirc), probably in a chemical solution, to seal the pores. (This is old info., but probably still correct.) If properly cared for, these plates might last a millennium or more...

Using alcohol to clean other surfaces, especially on older machines, (particularly those made before WW II) seems quite risky, especially finished wood! Try a tiny, inconspicuous spot, first.

}Hello all -
}
}First, about using alcohol for cleaning the surface of Curtas. I have }never experienced any problems with this - but I have always used what

Hello all - I've been enjoying the various postings about cleaning a gummed up calculator. Working around auto service centers gave me a hought - every mechanic who works on brakes seems to use a spray solvent designed for brakes: I'm going from memory but believe it's called "BrakeKleen" or something like that. It's a standard parts-house item. Also, there's a spray solvent that's used to clean tire innertubes. Both are extremely potent solvents. I would be almost 100 % certain that either one would damage or melt any plastic parts - and possibly paint. I haven't tried either one so it's simply a suggestion - something to try if all else fails.

Recently I purchased a late model Type II - it's within 600 numbers of another that I have owned for quite a while so I put them both on the kitchen table just for a comparison. Something was different but I couldn't figure out what it was. Finally, after a good bit of inspection, I realized what it was. On the one I had owned for a while, the "CURTA" above the sliders, the decimal position numbers from 1 to 15 at the bottom lip of the carriage, the slider decimal position numbers from 1 to 11 below the sliders, the arrow above the slider at position 1, AND the up/down arrow next to the reversing lever were all a very bright shiny silver. On the new Curta, these same numbers were dull and hazy - and it made a big difference in the overall appearance. Further checking made it clear that the silver I was seeing was not paint, but the metal underneath - and all of this had been engraved rather deeply after the various parts were finished. And the new machine had simply oxidized and lost the silver "shine". Always one to jump in with both feet, I carefully traced the "CURTA" using a dental pick and moderate pressure - and a good bit of care along with a steady hand. The result was amazing - the "CURTA" on both machines now looked virtually identical. So, I slowly went over all of the rest of the numbers (and arrows) that I mentioned. Now, it's very hard to see any difference between the two Curtas. It might be worth a try.

New topic. Recently, I've seen two different Curtas, both Type IIs, that had very bad damage to the upper or lower tracks for the decimal pointers, or both. It's a bit puzzling because I wouldn't really think they would be moved that much - but they obviously were. So I checked on several of my machines and found that the pointers didn't really move very freely on them, either. The pointers consist of a 1.0 mm steel ball and a tiny compression spring. I would assume that over the years, due to lack of lubrication, the steel balls either froze or wore down - and were scraping and scratching the black finished track. So, I took the smallest jewelers screwdriver and put a tiny drop of the Swiss clock oil I've mentioned down next to each pointer and then worked the pointers back and forth. Afterward, I used a couple Q-tips (you'll probably all think I own stock in Johnson & Johnson - I just find Q-tips very usable) to remove any excess oil from the tracks. I don't think this is anything that would have to be done frequently - but with many Curtas, it may never have been done since the Curta originally left the factory, so it doesn't seem like a bad idea to prevent damage little by little.

One thing that's always intrigued me about Curta (or really, Contina - the company's name) is that they were obviously small enough to change things as they desired. I have two very early sales brochures which show pictures of their two factories (I believe there were only two), one in Mauren and one in Eschen - and they look like rural, two- or three-story country inns. When you consider the estimated total of 140,000 units, Type I and II combined, and a 23-year production life, that only comes to about 24 per ay - if you assume that production was constant, which in all probability it was not.

As an example of this flexibility - I've identified 5 distinct, separate styles of Type I canisters:

1. original; high gloss with distinct shoulder on lid where the side becomes the top; additional wording besides "CURTA" and "<-----OPEN" ENGRAVED in the metal; normal or counter-clockwise opening;
2. same shape; high gloss; only "CURTA" and "<-----OPEN" engraved; reverse thread;
3. same shape; dull finish; attached labels rather than engraving; 4. domed shape on lid, like every Type II metal canister I've seen; dull finish; attached labels;
5. plastic canister - same basic shape as Type II plastic canister.

Type II canisters are more consistent in my experience: early is high gloss and engraved; later is same shape but dull finish and labels, and then last is plastic.

There seem to be new members all the time. Perhaps it's a good time to mention that there is a man in the U.S. who still repairs Curtas, for those of you who aren't aware of it. I offer the information strictly as a service to Curta owners and have no business connection with him. Please email me if you'd like his email address. Also, he is very low on parts and has mentioned that he'd be interested in buying any "broken" Curtas just for the parts.

Thanks to Bob DeCesaris and Rick Fisch, here is the information about the earliest type of case. The very early cases, typically for Type Is with round pegs for sliders, are a very very dull finish much like the Curta itself. There are 3 lines of engraving: [Universal-Rechenmaschine] on the top line, ["CURTA"] in the center, and in quotations, and [System Curt Herzstark] on the bottom line and in upper and lower case. Rick mentioned that the case also includes the machine serial number on the bottom, something that I don't believe is on any subsequent cases.

Thanks to Skip Solberg, the leather carrying case really should be included. I think most collectors are aware that there were two types of leather cases for each machine type - at least there are two types listed in an early sales brochure from The Curta Company in Van Nuys California. One holds the Curta inside the canister; the other holds just the Curta itself. Both could be included but certainly the latter one fits the description.

So, I believe that makes 7 variations for Type I machines. There's a possibility that the "System Curt Herzstark" appeared in both Upper/Lower case and also just in Upper case.

One final item. In two early brochures that I have, ones which happen to be in French and which are very similar but not identical, one shows the CURTA; the other one shows the same machine but it's labeled CONTINA, the name of the manufacturing company. Has anyone actually seen a CONTINA - or for that matter, a LILLIPUT?

That is the name that appears on the machine in a drawing in some other early literature and is apparently the name that Herzstark used prior to WW II. The basic design of the machine had been completed, or at least roughed out, prior to the war - but it's my understanding that it had not reached the point of any sort of preliminary production.

}My guess the body is an alloy of aluminum. The spiral shafts and }gears are probably steel.

The steel parts are probably several different alloys, according to their mechanical requirements and (possibly) whether they might need to be easy to machine. Stainless steel (corrosion-resistant) might well be used for many parts. Springs are a spring alloy, perhaps the same as music wire.

Much of the exterior is, indeed, probably Al alloy, perhaps anodized and dyed black. The bearings for the setting shafts (a delightful touch!) are porous bronze. The main shaft also probably runs in porous bronze bearings. (The pores hold oil.) Some parts might be made by powder metallurgy, but unless one sees them, or they're described as as such, one can't tell.

Detent and interlock balls are likely to be a steel alloy with a good amount of chromium.

I really don't know much about metallurgy so can't comment. I do know that all the metals seem to be impervious to rusting - or to oxidizing. One interesting sidelight. The first Curta that I bought, in 1993, included several original brochures bearing the stamp of a man in Pennsylvania who sold surveying equipment. I checked and sure enough he was still around. Among other things, he had sold his remaining stock of brand new Curtas just a year or two earlier - to a dealer, of course, and for the original selling price ($125 for Type I and $165 for Type II). But one other thing that he mentioned is what is at least a little bit relevant to this discussion.

The Univ. of Penn. was doing a lot of archeological digs in the rain forests of South America. In the very early '70s, when handheld electronic calculators became available, the purchased a number of them to be used in the rain forests - I assume that they needed to make the same calculations that regular land surveyors make. They quickly found that they wouldn't hold up in the humidity - mold, rot, fungus, all those good things would set in rapidly and short out the circuitry. But the Curtas they had were unaffected. So they junked all the electronic units and purchased a bunch more Curtas from him. He said they might still be in the University ystem - I checked and got several very nice responses but they were all gone: either sold for peanuts as surplus, or simply junked. The ability to hold up without reacting to the constant humidity would seem to confirm both the use of aluminum and high-quality ferrous alloys.

I also know from a collector friend in Germany that Curt Herzstark was a near-fanatic about quality. He met Herzstark in 1988, I believe, not too long before his death, and spent most of a day just talking to him. Herzstark apparently fought vigorously with the company officials at Contina over the use of plastic instead of metal in any part of the Curta. Interestingly, the broken clearing levers that turn up are invariably metal ones, not the later plastic ones which are flexible enough to withstand a jarring drop.

I stumbled onto a listing on eBay for a man in Colorado who has a fascinating assortment of super-strong magnets that he sells - like kids of all ages, magnets fascinate me. I got some from him just to play with - I'll see if I can tell whether any of the exterior parts on the Curta are ferrous - but I don't think that they are. If it's of interest, the URL for this website is http://www.wondermagnet.com/dev/index.html . The man is Dan Bartmann - nice guy; my order was delayed due to extra demand so he included some extras at no charge.

Concerning the metals once again. I saw one Curta for sale - probably assembled from leftover parts by the man in Liechtenstein who retired from Curta and still works on the machines. It appeared to look like a regular Curta except for two things - there was no "knurling" on either the upper or lower part, and there was no black finish on any part of the machine. The visible parts all had a color that appeared to be more aluminum than steel to me, but that's hard to tell also. I don't know if you're able to get JPGs off my emails, Nicholas - let me know (anyone else, too, and I'll send a JPG of that one: I'm fairly certain that I have it). Again, I assume that among all the other parts he reportedly acquired from Contina when Curta production stopped, he had a supply of parts from different stages of manufacturing and finishing. Last thought. The president/chairman of the company that absorbed Contina has two Curtas - a Type I and Type II - which are gold-plated. He loaned them to an exhibition that was held in Liechtenstein several years ago by a Swiss collector's group - I was never able to find out any more about that exhibition or the gold Curtas.

Hello all - I guess I should become a "snake oil" salesman: I apparently convinced Ernie Jorgenson that I knew how to repair Curtas. I wish I did - but I don't. I've ventured into a couple minor things with several of mine, but nothing that would qualify as a "repair" and I really don't think I would attempt that. The mechanism is simply too detailed and too miniaturized for a rank amateur to tackle.

The man who DOES have this knowledge, and who has worked on several of my machines, is Jack Christensen in Northbrook Illinois. This information has been sent to the list in the past but there are always new subscribers. Jack is sincerely interested in keeping as many Curtas humming away as possible. I think he does most of the work in his spare time so he can get overloaded on occasion. For anyone who needs to contact him about repair work, or simply a thorough cleaning and lubing, his email is: chronos@northshore.net

Curta News Letters

From Erez Kaplan's Machines List server

INTRODUCTION

CALCULATION OF ROOTS

SQUARE ROOTS (Hermann's Method)

CUBE ROOTS

Concise Instructions for Curta Multiply and Divide:

For Multiplying:

For Dividing: